A well-written case study is one of the most effective pieces of content a marketing agency can publish. It demonstrates real results, builds credibility, and shows prospective clients exactly what working with you looks like. But every time we sit down to write one at Choco Media, we face the same tension: the most compelling details — the specific numbers, the client name, the competitive context — are often the ones we can’t share. Agency case study writing is an exercise in making honesty and discretion coexist.
This post is about how we navigate that. It’s aimed at agency owners, strategists, and account leads who want to publish client work that is genuinely useful to readers without breaching trust or NDA terms. What follows is our working process — the questions we ask, the anonymisation techniques we use, and the structural choices that make anonymised case studies read as credible rather than vague.
By the end you’ll have a repeatable framework you can apply to your next client win, whether the client is happy to be named or categorically refuses to be mentioned at all.
Start With the Permission Conversation, Not the Writing
The biggest mistake agencies make with case studies is treating permission as an afterthought. They write a polished piece, then go to the client for sign-off, and discover the client is uncomfortable with three things they’ve already built the narrative around. That’s a painful rewrite.
We have a short permission conversation at the start, not the end. It usually happens during onboarding or at the point where results are materialising. We ask three questions:
- Can we reference your company by name, or would you prefer anonymised? Some clients are delighted to be named. Others have internal PR policies that make it impossible. Both are fine — we just need to know early.
- What categories of data can we publish? Traffic, revenue impact, conversion rates — some clients are comfortable with percentages but not absolutes. Others are happy with the sector average but not their specific number.
- Do you need final approval before we publish? Some clients want a sign-off round; others simply ask to be notified. Clarifying this avoids a last-minute delay when you’re ready to hit publish.
Get the answers in writing — even an email summary is enough. It protects the client relationship and protects you if there is ever a dispute about what was agreed.
Decide What the Story Is Actually About
Before you think about what you can and can’t share, decide what the story is. The strongest case studies are not documentation of what happened — they are arguments for a specific point of view. What does this piece of client work prove that your prospective clients need to believe?
The argument-first approach
In client work we’ve found that case studies written as arguments are far more persuasive than those written as chronologies. A chronology says: “In January we did X, in February we did Y, the results were Z.” An argument says: “When you fix this specific type of problem, this is what happens — and here’s the proof.”
The argument-first approach also makes anonymisation easier. Once you know the core claim you’re making, you can see which details are essential to that claim and which are just colour. Many of the details clients are uncomfortable sharing — their exact company name, their specific revenue figure — are often colour. The claim can stand without them.
- Identify the single insight the case study proves
- Map every detail to “essential to the argument” or “context / colour”
- Colour can almost always be replaced with an aggregate (“brands in this sector typically see…”) or a relative metric (“a 40% improvement over baseline”)
The Anonymisation Toolkit
Anonymising a case study is not the same as vagueing it out. Done well, an anonymised case study reads as specific and credible. Done badly, it reads like it might be made up. The difference is in the techniques you use.
Sector-level framing
Instead of naming the client, name the type of client. “A Scandinavian e-commerce brand selling premium outdoor gear” is more useful to the reader than just “a client.” It lets them pattern-match to their own situation without requiring you to reveal the company name. The more specific the sector framing, the better — vague sector labels like “a retail client” lose the reader immediately.
Relative rather than absolute metrics
If the client won’t permit you to publish their conversion rate, publish the improvement. “Conversion rate increased by 34% over 90 days” is more useful to a prospect than the underlying number anyway — it tells them the magnitude of the impact without revealing the client’s baseline performance, which is usually the commercially sensitive part.
Process specificity as a substitute for outcome specificity
If outcomes are truly off the table, go deep on process instead. The reader wants to understand what you do and whether you know what you’re doing. A detailed, honest account of the methodology — the audit steps, the creative iterations, the decision frameworks — demonstrates competence just as effectively as a result, sometimes more so.
- What did you find in the initial audit?
- What hypothesis did you form?
- What did you test and in what sequence?
- What did you change based on early results?
- What did you learn that you didn’t expect?
Process specificity is also more defensible. Numbers can be questioned; a well-described process is harder to dismiss.
Composite cases
If you have done similar work for several clients in the same sector and none will agree to be named, consider writing a composite case study. Be transparent that it is a composite — “the following is drawn from three client engagements in this sector” — and you lose nothing in terms of credibility. The composite can be genuinely more useful to the reader than any individual case because it captures the pattern across multiple instances, not just one.
The goal is not to protect the agency’s interests at the client’s expense or vice versa — it’s to find the version of the story that is honest, useful to the reader, and comfortable for the client. Those three things are almost always compatible if you look for them.
The Structure That Works for Anonymised Cases
The classic case study structure — situation, complication, resolution, results — works well even when anonymised. The structure itself signals credibility. Here’s how we typically adapt it:
Situation
Describe the client’s context at the sector and challenge level. You don’t need the company name to write a useful situation section. “A B2B SaaS business with an established product but declining organic traffic, operating in a market where competitors had significantly increased their content investment” is a situation any prospect in that position will recognise.
Complication
What specific problem needed solving? This is where process specificity earns its keep. What did your audit reveal? What were the root causes? Be honest about what you found — including the things that weren’t working — because that’s what demonstrates genuine diagnostic skill.
Resolution
What did you do? Walk through the approach in enough detail that the reader understands both the what and the why. This is the section that separates competent agencies from ones that just report what happened. Our AI content creation work often goes here — the methodology we use and why we made specific creative or technical decisions.
Results
Publish what you can. A relative improvement over a period of time is usually publishable even when the absolute number isn’t. If results are truly off-limits, end with what the client was able to do differently as a result of the work — “the team now had a content workflow that required 60% fewer hours per week” is a result without being a metric.
Writing the Case Study: Language and Tone
Anonymised case studies fail in one of two ways. Either they over-hedge — “we worked with a client who may or may not have seen some results” — or they over-claim in compensation for the lack of specificity. Both undermine credibility.
The language principles we use:
- Describe what happened, not what it was like. “Traffic to target pages increased 40% over 12 weeks” rather than “we achieved significant growth.”
- Use precise process language even when outcomes are hedged. “We ran three rounds of A/B testing on the headline before settling on a version” is specific and credible.
- Acknowledge complexity. “The first iteration underperformed, which led us to rethink the audience segmentation” reads as honest. Most clients appreciate this — it shows the agency is reflective, not just promotional.
- Write in plain language. Case studies are not the place for industry jargon. If your prospective client can’t understand what you did, they can’t assess whether it’s relevant to them.
A useful exercise: after drafting, read through and mark every sentence where you’re asserting something without evidence. Either cut those sentences or replace them with a specific detail that supports the assertion.
The Client Review Process
Even when the client has said they don’t need sign-off, we send a preview before publishing. It’s a five-minute gesture that prevents surprises. Our standard note is brief: “We’re planning to publish this case study next week. We’ve anonymised [X, Y, Z] as agreed — let us know if anything needs adjusting.”
In our experience this preview step almost never triggers significant changes. But the two or three times it has, we’ve been grateful for it — usually the client has flagged something we hadn’t thought to anonymise, or there’s a nuance in the situation description that isn’t quite right. Better to know before publication than after.
What to do when a client changes their mind
Occasionally a client who agreed to be named will change their mind after the case study is written. This most commonly happens when there’s been a change of personnel, or when the business is going through a sensitive period such as a funding round or acquisition.
- Don’t argue. Retrofit anonymisation using the techniques above. It’s usually a 30-minute job if the case study was well-structured.
- If the piece is already live, update it quickly and without drama. Add a brief note that the case study has been anonymised at the client’s request.
- Revisit the permission conversation with any future client at the end of engagement, not just the beginning. Circumstances change.
How We Use AI in Our Case Study Process
We use AI at two points in the case study workflow. The first is in the initial structuring pass — we feed in our raw notes from client calls, performance data exports, and internal strategy documents, and ask the model to identify the core narrative thread. This saves significant time and often surfaces angles we hadn’t considered.
The second is in the anonymisation pass itself. Once we have a draft, we run it through a check specifically designed to flag any detail that could identify the client — location specifics, unusual industry terminology, timeframes that coincide with known events. The model catches things that are easy to miss when you’re close to the material.
What AI doesn’t do is replace editorial judgement. The decision about what to include, how to frame a complication honestly, or how to describe a result without over-claiming — those remain human calls. We’ve found this split works well: AI for pattern recognition and completeness checking, humans for judgement and tone. If you want to understand more about how we structure this division, our post on AI for client communication covers the broader principle.
Publishing and Promotion
Once published, treat the case study as a content asset, not just a page on your website. The work we put into making a case study specific and honest pays off most when it’s placed in front of people who are actively considering the type of work it demonstrates.
- Link to it from relevant service pages — the case study is evidence for the service, so put the evidence where the prospect is looking.
- Send it to prospects who are in the consideration stage for similar work. A well-targeted case study is more persuasive than a general capability overview.
- Use excerpts in proposals. The specific details that earned credibility in the case study will do the same work when embedded in a proposal document.
- Revisit and update annually. Results accumulate; the story often gets better over time.
For a deeper look at how we handle the broader proposal process, including how case studies function within it, our post on agency proposal structure walks through the full sequence we’ve landed on after 50+ pitches.
The Ethics of the Anonymised Case Study
A question worth sitting with: is an anonymised case study still honest? Our answer is yes, provided two conditions are met. First, the anonymisation should not change the substance of what is being claimed — you shouldn’t describe a process as yours when it was largely the client’s. Second, you should be transparent about the anonymisation rather than presenting a fictional company as if it were real.
Beyond those two conditions, anonymisation is not a compromise of honesty — it’s a condition of the professional relationship. Clients share commercially sensitive information with us precisely because they trust us to handle it appropriately. Publishing that information without consent would be a breach of that trust regardless of how compelling the story would be.
If you’re ready to start building case studies that work — and you want support with the content strategy around them — get in touch and we can talk through what that looks like for your agency.